A Novel Risk Analysis Approach of Casing under Complex Conditions Using Copulas
Lei Zhao,Yifei Yan,Xiangzhen Yan
DOI: https://doi.org/10.1016/j.petrol.2020.106955
IF: 5.168
2020-01-01
Journal of Petroleum Science and Engineering
Abstract:Failure accidents of casings, designed obeying traditional design standards that have been amended for many times, are still happening nowadays, and existing reliability analysis methods mostly have not considered the dependence structures of variables, it's badly necessary to seek other methods to compensate these two problems and defects. This study proposes a copula-based risk analysis framework of casing in a casing-formation composite system (CFCS). Firstly, a mechanical analysis model of casing under CFCS is established and verified by the simulation method. Secondly, random data sets of the formation parameters are obtained by using logging data, meanwhile, the acquisition methods of statistical data of casing are also provided. Thirdly, the best uncertainty parameters in the given five copulas are inferred by using Bayesian-based Markov chain Monte Carlo (MCMC) likelihood algorithm, which are compared with those from other algorithms; the optimal copula is determined by the Bayesian-based MCMC-estimated method and the squared Euclidean distance (SED), respectively; the posterior joint probability isolines of the copulas are depicted, and isolines from the optimal copula are compared with that of the empirical copula method; the correlation coefficients of variables in CFCS, relative to performance function value, are sorted in magnitude. Finally, a division method of safety level is adopted, and based on the optimal copula, the risk analysis of casing is conducted with considering the bivariate dependence. The proposed framework can provide unique insights into the risk analysis of casing in CFCS with correlations considered and can also be programmed to use in risk management and decision-making systems of petroleum engineering.